People with Alzheimer disease may be less likely to develop cancer, and people with cancer may have a lower risk of AD, according to a large population-based epidemiological study published online this week in Neurology. The work, from Catherine Roe and colleagues at Washington University in St. Louis, Missouri, puts legs under the idea that the molecular pathways underlying cancer and neurodegeneration could be related—a concept that that has been percolating through the field for some time.

The new study follows up on previous work from the same researchers (Roe et al., 2005) that suggested an inverse relationship between the two diseases. This time, Roe and colleagues looked at the question in a group of 3,020 people age 65 or older who had enrolled in a long-term cardiovascular health study. The investigators asked whether time to a first hospitalization for cancer was related to the prevalence of dementia, and conversely, whether the chances of dementia were changed in people with a history of cancer. The subjects were followed for an average of 5.4 years for dementia and 8.3 for cancer.

The results showed that people with a diagnosis of AD at baseline had a 60 percent reduced risk of cancer compared to subjects without AD. The association was specific for AD, as it was not seen with vascular dementia. Likewise, people with a history of cancer had a 30 percent lower risk of developing AD during follow-up.

In an accompanying editorial, David Bennett and Sue Leurgans of Rush University in Chicago point out some caveats, the most important of which is the possible effect of mortality on the observed associations. Nonetheless, they write, the uncovered associations are worthy of further exploration.

Paul Coleman of the University of Rochester, New York, wrote in an e-mail to ARF that he agrees with the cautions of Bennett and Leurgans, adding that he has previously looked at comorbidity of AD and cancer using an extensive database at the Marshfield clinic in Wisconsin. That dataset is a remarkable one, he wrote, with extensive records on a relatively stable cohort of tens of thousands of people. “In a nutshell, these data showed no relationship between the two (once age was factored out). But I was unsatisfied with the ascertainment of the cancer cases (all lumped together) so this result has never been published,” Coleman wrote.

Roe’s results, pending replication, would support the idea that there may be a common molecular mechanism linking cancer and AD. Independent studies have unearthed evidence for a similar link between neurodegeneration and cancer in Parkinson disease (Driver et al., 2007; Inzelberg et al., 2007). The common thread could be pathways implicated in cell death and survival, or epigenetic regulation (see ARF related news story on Mastroeni et al., 2009). Alternatively, perhaps chemotherapy has an unexpected upside for neurons—a recent study suggests that women who had chemotherapy for breast cancer were at a reduced risk of AD (Du et al., 2009). Whatever the reason, the concept of AD and cancer as opposite sides of the coin of age-related diseases could lead researchers to new insights into neurodegeneration. This coming April, the Ipsen Foundation will hold a conference entitled: Two Faces of Evil: Cancer and Neurodegeneration in Paris, France.—Pat McCaffrey


  1. The article by Roe et al. is a strong contribution to the literature of two fields—cancer and AD. But while the field will benefit from having access to the data and the analyses reported, the article and the accompanying editorial bring up two questions in my mind.

    The first is a solely theoretical one. In their accompanying tables, the authors cite the ApoE profiles of the two groups (those getting cancer and those getting dementia) but unfortunately do not comment on the data itself. This is frustrating, because the strong correlation between carrying one or two ApoE4 alleles and elevated AD risk means a potential insight into mechanism has slipped through their fingers. The sample size is large enough that they should replicate the often-observed AD/ApoE4 connection in their dementia population. But then, according to their hypothesis, the cancer data should go the other way, i.e., ApoE4 genotype should be protective. The 4/4 numbers are small, but seem adequate given that increased risk of AD for this group has been estimated to be above 10-fold. I don't see this effect in the paper’s cancer table, although I do see it (in the expected direction) in the dementia table. If my cursory impression is true, and there is no ApoE4-cancer linkage, it would hint that the cancer/AD connection is working off a parallel biological mechanism that is independent of the ApoE mechanism. And since it is widely believed that ApoE cannot account for all cases of sporadic AD, this could be a very important insight.

    My second comment is both theoretical and personal. On behalf of the many laboratories (Herrup, Arendt, Davies, Lamb, Neve, Copani, Chun, Slack, Park, Potter, Greene, Benes, Dawson, Julien, Bowser, Sharp, Rakic, Smith and Perry, and probably others) that have labored hard over the past 15 years to repeatedly demonstrate the linkage between loss of cell cycle control in adult nerve cells and the onset of neurodegeneration in AD and other diseases, I wish to express our concern that neither the authors of the study nor the writers of the editorial mentioned this rapidly growing body of work and how it relates to their findings. In bringing up Pin1, which is more closely associated with DNA damage than with cell cycle regulation, they also ignore the excellent work on the association between mutations in DNA repair enzymes and developmental nerve cell death (e.g., ATM, ATR, Mre11, Nbs1, etc.). I mention this partly out of personal pride, but mostly because it raises a potentially important theoretical question about mechanism. In every case, DNA damage and/or loss of nerve cell cycle control is associated with increased nerve cell death. Therefore, on first principles we might expect that the correlation between AD and cancer (also accompanied by DNA damage and a loss of cell cycle control) would be positive, not negative as reported here. Even putting aside my obvious personal interest, I suggest that the lack of discussion of this point by the authors is an opportunity missed.

  2. Comment by Hyoung-gon Lee, Xiongwei Zhu, Xinglong Wang, Rudy J. Castellani, George Perry, and Mark A. Smith

    Cancer, Alzheimer Disease, and Cardiovascular Disease: Associations and Disassociations—Clues to Etiology?
    There are numerous pathological similarities between cancer and Alzheimer disease (AD), including, but not limited to, loss of cell cycle control (McShea et al., 1997), DNA instability/replication (Bajic et al., 2008; Spremo-Potparevic et al., 2008; Zhu et al., 2008; Bajic et al., 2009), abnormal proliferative signal transduction pathways (McShea et al., 1999; Perry et al., 1999), and avoidance of apoptosis (Raina et al., 2001; Zhu et al., 2006). Additionally, it is notable that numerous features of AD follow the oncogenic stimulation of neurons either in vitro (McShea et al., 2007) or in vivo (Lee et al., 2009a). Given such striking similarities, it is perhaps surprising that AD is associated with a reduced risk of cancer and, likewise, that cancer was associated with a reduced risk of AD (Roe et al., 2010). However, since it is clear that the etiologies of both AD and cancer are likely multifactorial and involve a series of “hits” (Zhu et al., 2001; Zhu et al., 2004; Zhu et al., 2007), one possible explanation is that the risk factors predisposing to cancer are protective against AD and, conversely, that the risk factors for AD are protective against cancer. In this regard, it is worth noting that the etiopathogenesis of cancer involves stem cells and perhaps other somatic cell types that have proliferative potential, whereas AD affects neurons that are terminally differentiated. Supporting such an association, cardiomyopathy and heart disease are associated with an increased risk of AD and are conditions associated with another terminally differentiated cell, the cardiomyocyte (Lee et al., 2009b). Obviously, such associations and disassociations might provide clues to etiology.


    . The X-chromosome instability phenotype in Alzheimer's disease: a clinical sign of accelerating aging?. Med Hypotheses. 2009 Dec;73(6):917-20. PubMed.

    . Is the time dimension of the cell cycle re-entry in AD regulated by centromere cohesion dynamics?. Biosci Hypotheses. 2008;1(3):156-161. PubMed.

    . The neuronal expression of MYC causes a neurodegenerative phenotype in a novel transgenic mouse. Am J Pathol. 2009 Mar;174(3):891-7. PubMed.

    . Cell cycle re-entry and mitochondrial defects in myc-mediated hypertrophic cardiomyopathy and heart failure. PLoS One. 2009;4(9):e7172. PubMed.

    . Abnormal expression of the cell cycle regulators P16 and CDK4 in Alzheimer's disease. Am J Pathol. 1997 Jun;150(6):1933-9. PubMed.

    . Neuronal cell cycle re-entry mediates Alzheimer disease-type changes. Biochim Biophys Acta. 2007 Apr;1772(4):467-72. PubMed.

    . Signal transduction abnormalities in Alzheimer's disease: evidence of a pathogenic stimuli. Brain Res. 1999 Jan 9;815(2):237-42. PubMed.

    . Activation of neuronal extracellular receptor kinase (ERK) in Alzheimer disease links oxidative stress to abnormal phosphorylation. Neuroreport. 1999 Aug 2;10(11):2411-5. PubMed.

    . Abortive apoptosis in Alzheimer's disease. Acta Neuropathol. 2001 Apr;101(4):305-10. PubMed.

    . Cancer linked to Alzheimer disease but not vascular dementia. Neurology. 2010 Jan 12;74(2):106-12. Epub 2009 Dec 23 PubMed.

    . Premature centromere division of the X chromosome in neurons in Alzheimer's disease. J Neurochem. 2008 Sep;106(5):2218-23. PubMed.

    . Differential activation of neuronal ERK, JNK/SAPK and p38 in Alzheimer disease: the 'two hit' hypothesis. Mech Ageing Dev. 2001 Dec;123(1):39-46. PubMed.

    . Alzheimer disease, the two-hit hypothesis: an update. Biochim Biophys Acta. 2007 Apr;1772(4):494-502. PubMed.

    . Alzheimer's disease: the two-hit hypothesis. Lancet Neurol. 2004 Apr;3(4):219-26. PubMed.

    . Apoptosis in Alzheimer disease: a mathematical improbability. Curr Alzheimer Res. 2006 Sep;3(4):393-6. PubMed.

    . Neuronal binucleation in Alzheimer disease hippocampus. Neuropathol Appl Neurobiol. 2008 Aug;34(4):457-65. PubMed.

    View all comments by Xiongwei Zhu
  3. One molecular mechanism that could explain this is sodium. In the Hypothesis Factory (1) I explain how repeated osmotic swelling of the brain resulting from hyponatremia could be a root cause of Alzheimer’s. It is widely believed that a high-salt diet is somehow responsible for a higher rate of stomach cancer. This may explain why the Japanese have a higher rate of stomach cancer coincident with a lower rate of Alzheimer’s.

    1. Could Hyponatremia Be the Root Cause of Alzheimer's?

Make a Comment

To make a comment you must login or register.


News Citations

  1. Twin Study Suggests Epigenetic Differences in AD

Paper Citations

  1. . Alzheimer disease and cancer. Neurology. 2005 Mar 8;64(5):895-8. PubMed.
  2. . Are Parkinson disease patients protected from some but not all cancers?. Neurology. 2007 Oct 9;69(15):1542-50. PubMed.
  3. . Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer's disease. PLoS One. 2009;4(8):e6617. PubMed.
  4. . Relationship between chemotherapy use and cognitive impairments in older women with breast cancer: findings from a large population-based cohort. Am J Clin Oncol. 2010 Dec;33(6):533-43. PubMed.

External Citations

  1. Driver et al., 2007

Further Reading

Primary Papers

  1. . Cancer linked to Alzheimer disease but not vascular dementia. Neurology. 2010 Jan 12;74(2):106-12. Epub 2009 Dec 23 PubMed.
  2. . Is there a link between cancer and Alzheimer disease?. Neurology. 2010 Jan 12;74(2):100-1. PubMed.